The catalytic activity of transition-metal-based atomically dispersed catalysts is closely related to the spin states. Manipulating the spin state of metal active centers could directly adjust the d orbital occupancy and optimize the adsorption behavior and electron transfer of the intermediates and transition metals, which would enhance the catalytic activity. We summarize the means of manipulating spin states and the spin-related catalytic descriptors. In future work, we will build a quantifiable and accurate prediction intelligent model through artificial intelligence (AI) and machine learning tools. Furthermore, we will develop new spin regulation methods to carry out the directional regulation of atomically dispersed catalysts through this model, providing new insight into the rational design of transition-metal-based atomically dispersed catalysts through spin manipulation.